Hypoglycemia is a feared complication of diabetes therapy, with consequences ranging from inconvenience to death. In the ACCORD cohort, severe symptomatic hypoglycemia was associated with an increased risk for mortality. The purpose of this project is to use the ACCORD cohort to identify biomarkers that would predict severe hypoglycemia, with the long-range goal of individualizing diabetes therapy based on selection of the safest approach for each patient. In this project we will test the overall hypothesis that two biomarkers of insulin deficiency - reduced fasting serum C-peptide concentrations and detectable levels of antibodies associated with islet autoimmunity - are strongly associated with severe hypoglycemia and death in the ACCORD cohort. Our specific aims are: Aim 1a: To measure baseline fasting serum C-peptide at baseline in subjects who died and experienced severe hypoglycemia during the ACCORD trial and compare to baseline values measured in matched ACCORD subjects who did not experience death and severe hypoglycemia during the trial. Aim 1b: To measure baseline autoantibodies to GAD 65 (GAD), intracytoplasmic domain of the tyrosine phosphatase-like protein IA-2 (IA-2A) and insulin (IAA) in subjects who died and experienced severe hypoglycemia during the ACCORD trial and compare to baseline values measured in matched ACCORD subjects who did not experience death and severe hypoglycemia during the trial. Aim 2a: To measure fasting serum C-peptide at baseline in subjects in the intensive arm who experienced severe hypoglycemia despite failing to achieve a hemoglobin A1c < 6.0% at any visit prior to study completion or death and compare to values measured in matched ACCORD subjects in the intensive arm who achieved target glycemia on at least one occasion and completed the trial without severe hypoglycemia. Aim 2b: To measure GAD, IA-2A, and IAA at baseline in subjects in the intensive arm who experienced severe hypoglycemia despite failing to achieve a hemoglobin A1c < 6.0% at any visit prior to study completion or death and compare to values measured in matched ACCORD subjects in the intensive arm who achieved target glycemia on at least one occasion and completed the trial without severe hypoglycemia.